Abstract: The prediction of ship main engine fuel consumption is the basis and premise of ship energy efficiency optimization,and the analysis of ship fuel consumption prediction results under different sailing areas can more effectively improve the prediction performance of the fuel consumption model.This paper selects five voyage segments as the experimental objects according to the sailing areas and establish the fuel consumption model,analyze the factors affecting the main engine fuel consumption select the main engine speed,wind speed,wind direction,etc.as the input variables of the model,and the instantaneous fuel consumption of the main engine and the voyage speed as the output variables,and use the back-propagation neural network to make predictions on the model.The experimental results show that the prediction errors for fuel consumption and speed are no more than 2.5% and 1.8%,respectively;however,in segment 2and segment 3,where the wind varies more steadily,the corresponding prediction errors are lower than those in other segments.The final experimental results show that the prediction accuracy of the fuel consumption model is affected by the degree of wind variation,but the prediction performance in different segments can meet the requirements for subsequent energy efficiency optimization.